26 December 2009

Google Races To Speed Up The Web

Business Week



Google is in a really big hurry to make the Web experience faster.

That became abundantly clear in 2009, when the company unveiled a steady stream of projects, products and enhancements focused on increasing Internet speed.

The initiatives varied widely in scope and focus. They included organic Google search engine optimization tools for webmasters to improve Web site performance. Others were speed-focused products like the Go programming language and Chrome OS. Google also pushed proposals to revamp aging Internet protocols and lobbied governments for broadband improvements.

The common denominator across the various heterogeneous efforts is an urgent desire to speed up the Web. This is rooted in Google's deeply held belief that a faster Internet is good for everyone, drives up online usage all around and boosts the company's business.

"The benefit of speed improvements is very substantial for the overall Internet. In the end, as the Internet gets faster, we benefit because people will use it more, which hopefully means they'll do more searches, which gets us more revenue," said Urs Hölzle, a Google fellow and senior vice president of operations.

At first pass, it's hard to argue with this reasoning, particularly when many of the Google efforts are free and available as open-source software, which anyone can adopt, modify and use.

Plus, Web latency remains a chronic, thorny problem with many improvement opportunities, and Google has the financial and talent resources needed to lead the way and tackle the bottlenecks.

"Google can afford to develop these projects in a way that's going to be strategic," said Sheri McLeish, a Forrester Research analyst. "Google has a luxury most companies don't. They're running a very successful business and what we're seeing is what that success affords them."

Still, as Google advances aggressively with its speed initiatives, it also indirectly increases its already considerable power and influence over people's and organization's Web activities, raising so-called "Goog-zilla" concerns.

For example, in early December, as part of this overall speed initiative, Google launched a new system to resolve DNS (domain name system) queries. It said its product improves on existing DNS resolver technology with faster, more efficient caching and additional security safeguards.

The thing is, those who switch to Google Public DNS will let the company know which Web sites they're visiting. This adds to the already massive amount of user data Google has access to via its search engine, online services and applications.

"Google definitely sees itself as being in a position to influence the future direction of the Web. From a business strategy perspective, it's smart of them, because clearly they're a major beneficiary of Web traffic," said Hadley Reynolds, an IDC analyst.

"However, many people are very concerned that when you add up this whole picture there can be some deep nefarious purpose behind all of these moves. That's a legitimate concern," Reynolds added.

Reynolds notes similarities between Google's rise to power on the Web using its leadership in search and the way in which Microsoft used its dominance of the PC OS to expand into many other areas of consumer and enterprise computing.

"When you add up all the elements, it amounts to essentially being a Windows kind of dominant player in the Web environment, creating the same kind of presence Microsoft has had on the PC and then the enterprise. Google clearly has a level of ambition to do the same kind of thing on the Web," Reynolds said.

"At this stage, no one in the Web environment has a similar platform from which to compete against Google," he added.

Hölzle said Google understands privacy concerns around its various products and initiatives, which is why the company strives to be clear and transparent in its policies related to user data collection, analysis and retention. After all, the company reasons, if people distrust Google, they will stop using its products and services.

For example, Google pledged that Google Public DNS will retain end-users' IP addresses for no longer than 48 hours, and that it will store for no longer than two weeks general data about the users' ISP and city. Google also promised not to use Google Public DNS traffic data to complement data it collects from users in its other services.

Nonetheless, as skeptics often point out, one must take Google at its word on faith, hoping it sticks to its promise not only now but also in the future, and that accidental data leaks or compromises don't occur.

Another concern is what some view as Google's sometimes short attention span, which causes the company to release early prototypes of often promising services that it later loses interest in and abandons. By Hölzle's estimation, Google has about 25 speed-focused projects in various degrees of completion.

"One of the bigger challenges for Google is delivering on its promises. It tends to make announcements very early in the development cycle to get people excited and interested, but then it seems to take a long time to move those products to production ready," McLeish said.

Hölzle said Google views its speed efforts as key to its business, and that its commitment to them is solid and long term, supported by the company's top executives. Google even has a virtual team focused on speed matters with members in a wide variety of product groups.

Hölzle also pointed out that Google's speed obsession isn't a fad that caught on this year. The company has been stressing the importance of speed for several years, and many efforts unveiled in 2009 had been in the works since 2008 and before.

One ambitious speed initiative is SPDY, pronounced "speedy," which aims to revamp the venerable HTTP protocol for Web content transport. SPDY, still in experimental phase, could be much faster and better able to take advantage of broadband, a sort of "HTTP 3.0," Hölzle said.

In this area of legacy Internet protocols, Google backs HTML 5.0, because the company believes its improvements greatly help to narrow the performance gap between Web applications and PC applications.

Changes at the Internet protocol and standards level require industry-wide goodwill, consensus and collaboration, which are sometimes tricky to generate due to competitive tensions among vendors. Hölzle trusts Google will succeed with its proposals, especially SPDY, for which it has high expectations.

"The Internet community is good about having technical arguments and making them based on technology merits. If we show why [our proposals are] good and we can give people an open-source implementation to demonstrate how it works and also measurements to see what the benefit is, I'm confident we can change standards because everyone will realize it's for everyone's benefit," he said.

Another major area of focus for Google is Web site design, where best practices are often ignored, even by major Web publishers, resulting on clunky and slow Web pages. Google has promoted good design principles and released webmaster tools like Page Speed. This open-source Firefox add-on analyzes a site's server and front-end design and, based on a set of best practices, generates recommendations.

Then there are the products created primarily because Google saw an opportunity to build a better mousetrap, like the Chrome browser, launched last year, and Chrome OS and Go programming language, both introduced this year. Likewise, many improvements across its product line are driven by speed considerations. "We see again and again that if you make something faster, the usage goes up almost immediately," he said.

It remains to be seen if, in fact, what makes Google SEO better and faster does indeed have similarly beneficial effects on the Internet community at large. With the velocity at which Google is pushing ahead its speed efforts, we'll likely find out soon enough.

No comments:

Post a Comment